Process Management • Process management deals with several issues: – what are the units of execution Processes & Threads – how are those units of execution represented in the OS – how is work scheduled in the CPU – what are possible execution states, and how does the system move from one to another Managing Concurrency in Computer Systems 1 2 The Process What’s in a Process? • Basic idea is the process: • A process consists of at least: – process is the unit of execution – the code for the running program – it’s the unit of scheduling – the data for the running program – it’s the dynamic (active) execution context (as opposed to a pro gram, – an execution stack tracing the state of procedure calls made which is static) – the Program Counter, indicating the next instruction • A process is sometimes called a job or a task or a sequential – a set of general -purpose registers with current values process. – a set of operating system resources (open files, connections to other • A sequential process is a program in execution; it defines the programs, etc.) sequential, instruction-at-a-time execution of a program. • The process contains all the state for a program in execution. 3 4 1 Process State Process State Changing • There may be several processes running the same program (e.g., an editor), but each is a distinct process with its own Terminated representation. New Ready Running • Each process has an execution state that indicates what it is currently doing, e.g.,: – ready: waiting to be assigned to the CPU Waiting – running: executing instructions on the CPU – waiting: waiting for an event, e.g., I/O completion • As a program executes, it moves from state to state Processes move from state to state as a result of actions they perform (e.g., system calls), OS actions (rescheduling), and external actions (interrupts) 5 6 Process Data Structures PCB • At any time, there are many processes in the system, each in its particular state. The PCB contains the process state • The OS must have data structures representing each process: this entire state of the process number data structure is called the PCB: process program counter – Process Control Block • The PCB contains all of the info about a process. stack pointer • The PCB is where the OS keeps all of a process’ hardware general -purpose registers execution state (PC, SP, registers) when the process is not running. memory management info username of owner queue pointers for state queues scheduling info (priority, etc.) accounting info 7 8 2 PCBs and Hardware State State Queues • When a process is running its Program Counter, stack pointer, • The OS maintains a collection of queues that represent the state of registers, etc., are loaded on the CPU (I.e., the processor hardware all processes in the system. registers contain the current values) • There is typically one queue for each state, e.g., ready, waiting for • When the OS stops running a process, it saves the current values I/O, etc. of those registers into the PCB for that process. • Each PCB is queued onto a state queue according to its current • When the OS is ready to start executing a new process, it loads the state. hardware registers from the values stored in that process’ PCB. • As a process changes state, its PCB is unlinked from one queue and linked • The process of switching the CPU from one process to another is onto another. called a context switch. Timesharing systems may do 100s or 1000s of context switches a second! 9 10 State Queues PCBs and State Queues PCB A PCB B PCB C Ready Queue Header • PCBs are data structures, dynamically allocated in OS memory. head ptr • When a process is created, a PCB is allocated to it, initialized, and tail ptr placed on the correct queue. • As the process computes, its PCB moves from queue to queue. Wait Queue Header • When the process is terminated, its PCB is deallocated. head ptr tail ptr PCB X PCB M There may be many wait queues, one for each type of wait (specific device, timer, message,…). 11 12 3 Cooperating Processes Processes and Threads • Processes can be independent or they can be cooperating to • A full process includes numerous things: accomplish a single job. – an address space (defining all the code and data pages) • Cooperating processes can be used: – OS resources and accounting information – to gain speedup by overlapping activities or performing work in – a “thread of control”, which defines where the process is parallel currently executing (basically, the PC and registers) – to better structure an application as a small set of cooperating processes • Creating a new process is costly, because of all of the – to share information between jobs structures (e.g., page tables) that must be allocated • Sometimes processes are structured as a pipeline where each • Communicating between processes is costly, because produces work for the next stage that consumes it, and so on. most communication goes through the OS 13 14 Parallel Programs “Lightweight” Processes • Suppose I want to build a parallel program to execute on a • What’s similar in these processes? multiprocessor, or a web server to handle multiple simultaneous – They all share the same code and data (address space) web requests. I need to: – they all share the same privileges – create several processes that can execute in parallel – they share almost everything in the process – cause each to map to the same address space (because they’re part of • What don’t they share? the same computation) – Each has its own PC, registers, and stack pointer – give each its starting address and initial parameters • Idea: why don’t we separate the idea of process (address space, – the OS will then schedule these processes, in parallel, on the various accounting, etc.) from that of the minimal “thread of control” (PC, processors SP, registers)? • Notice that there’s a lot of cost in creating these processes and possibly coordinating them. There’s also a lot of duplication, because they all share the same address space, protection, etc…… 15 16 4 How different OSs support threads Threads and Processes • Some newer operating systems (Mach, Chorus, NT) therefore = address space support two entities: – the process, which defines the address space and general process = thread attributes – the thread, which defines a sequential execution stream within a process • A thread is bound to a single process. For each process, however, example: MS/DOS example: Unix there may be many threads. • Threads are the unit of scheduling; processes are containers in which threads execute. example: Xerox Pilot example: Mach, OSF, Chorus, NT 17 18 Separation of Threads and Processes Kernel Threads • Separating threads and processes makes it easier to support multi- • Kernel threads still suffer from performance problems. threaded applications • Operations on kernel threads are slow because: • Concurrency (multi-threading) is useful for: – a thread operation still requires a kernel call – improving program structure – kernel threads may be overly general, in order to support needs of – handling concurrent events (e.g., web requests) different users, languages, etc. – building parallel programs – the kernel doesn’t trust the user, so there must be lots of checking on • So, multi-threading is useful even on a uniprocessor. kernel calls • But, threads, even when supported separately from processes in the kernel, are too slow. 19 20 5 User-Level Threads (Old) Example of thread performance • To make threads really fast, they should be implemented at the Ultrix Topaz FastThreads user level Fork 11320 1208 39 • A user-level thread is managed entirely by the run-time system (user-level code that is linked with your program). Signal/Wait 1846 229 52 • Each thread is represented simply by a PC, registers, stack and a little control block, managed in the user’s address space. (performance on a 3 MIPS processor, in microseconds) • Creating a new thread, switching between threads, and synchronizing between threads can all be done without kernel involvement. • Ultrix: 1 thread per address space • Topaz: multiple threads per address space • FastThreads: multiple user threads per address space 21 22 Example U-L Thread Interface t = thread_fork(initial context) create a new thread of control thread_stop() stop the calling thread, sometimes called thread_block thread_start(t) start the named thread thread_yield() voluntarily give up the processor thread_exit() terminate the calling thread, sometimes called thread_destroy. 23 6.